The Opportunity

Travel agents working within an international Hotel and Lodging chain could provide concierge service to high-end clientele. However, such well-planned travel itineraries and ideal accommodations remained beyond the reach of most of their customers. This company sought an AI product that could begin to deliver such services. 

Our Solution

Our Flask-based web application now offers tailored advice by gathering customer preferences from their profiles, search, and stay histories to generate travel itinerary recommendations. Our generative AI chat interface, alongside traditional machine learning recommenders, augments this labor-intensive process, resulting in an improved customer experience with minimal investment of company resources.

For this, we used the AWS stack including Elastic Kubernetes Services and containerized applications, making scaling a breeze. For the generative AI side, we used Postgres, along with Ada Embeddings from OpenAi. Our back-end was written in Python’s Flask library and the front-end was written in JavaScript/HTML/CSS.

Profiles and preferences were piped into Postgres via jobs in Sagemaker. CI/CD was handled with Jenkins and Artifactory. Internal API calls were written on top of a GraphQL layer that called REST services under the hood and offered real-time availability of rooms, along with information about pricing and points for budgeting purposes.